|
| Topic |
Issue |
| Final |
Q:
Is the Final cumulative?
A:
Definitely. Be sure in particular to review regression and probability,
two topics that are taught rather early in the semester!
|
| Final |
Q:
What chapters do we not need to worry about for the final?
A:
Excluded from the final:
ch. 7 on points and lines in the plane (but be sure you understand
how this material is used in the chapters on regression),
ch.25 on genetics,
any specific knowledge about
ch.19 (Sample Surveys) and ch. 22 (Unemployment, CPS) but be sure to
understand the connection between SRS and SE formulas!
No decision has been made yet what material, if any,
we guarantee not to be on the final.
|
| Final |
Q:
What will be on the Final?
Everything not excluded in the previous question is fair game.
Some specifics will be in one of the last lectures but here is a
list of particularly important topics:
-- ch.1-2 (Experiments):
Bias and confounding factors in experiments and surveys
-- ch.3-5 (Histograms):
density scale, class intervals, left/right tailed,
outliers, geometric meaning of
average and SD, SD vs SD+
-- Normal curve, standard units, change of scale, percentiles
-- ch.6 (Measurement Error):
All concepts, e.g., chance error, repeatability, outliers, bias
-- ch.8-12 (Regression):
SD line vs regression line (formulas), formula for r, residuals,
change of scale
-- Regression line as approximation of y-averages
of small intervals about x (ch.10, Fig.1)
-- Association vs causation, regression fallacy
-- r.m.s. error of the regression line (standard error of the estimate)
-- homo/heteroscedascity
-- Least squares and regression line
-- Probability: all of ch. 13-15, incl. the binomial formula
-- Ch.16-18 (Chance variability): box models and chance processes
as a means to draw samples
-- counting boxes and shortcut formula for SD of 2-ticket value boxes
-- SE for chance processes vs SDs for lists like boxes and samples
-- SE and expected value for sums, averages and percentages
-- square root laws for SEs
-- approximation of normal curve by the probab. histograms for
observed minus expected divided by appropriate SE
-- continuity correction
-- Ch.19-24 (Sampling):
-- Importance of independent or SRS sampling from a box, correction factor
-- Inferential statistics,
reasoning from the box to samples vs
inferences from the sample to the box
-- Connection between SE and confidence intervals
-- Ch.26-29 (Hypothesis Testing):
-- Null hypothesis vs alt. hypothesis, role of box model, SRS, ...
-- z-test, t-test, 2-sample z-test, when to use the t-distribution
-- chi2 tests for box composition and independence
-- left tailed chi2 tests for unlikely accuracy
-- P-value, level of significance, significant vs highly significant,
test statistics (z-statistic, t-statistic, ...)
-- one-tailed vs two-tailed tests, data snooping
-- real world relevance of hypothesis tests
|
| Final |
Q:
How do I start preparing for the final?
A:
a. Start out by reviewing all chapter summaries in the FPP book.
They are short and give you an overview of what's important.
b. Review the completely worked examples in the text book.
c. Solutions for one version of each quiz will be posted
You should review those.
|
| Exam |
Q:
Can we write on both sides of the index card?
A:
Yes.
|
| Exam |
Q:
What is the format of the exams?
Will there be multiple choice and fill-in-the-blanks questions?
Must we write down formulas?
Will table lookup pages be provided or can we use
our calculator?
A:
My intent is to reward those who understand more
of the course material than those who understand less.
Everything else is a consequence of the above:
You should have some judgement about subjects which are
of central importance and those that are peripheral.
One item that I'll be happy to clarify:
If you are asked to compute a number such as
the standard deviation of a list of, say, 5 specific numbers,
then you must provide the formula where those numbers are plugged in.
That will get you most of the credit whereas
just writing the result (which you might have gotten
by pushing a couple of buttons on your statistics calculator)
will give you very little credit.
And, yes, you are expected to know definitions such as those
of the SD of a list (how else could you write down
the formula with the values plugged in?)
or of a histogram or of an SD line so you'll be able
to graph them in case you are asked to do so.
And yes, some questions will be multiple choice and fill-in-the-blanks.
Table-lookups:
If you need to get numbers that are associated with
the normal curve then you must do the look-ups in
the z-scores table on page A104 of the book
(it will be attached to your exam) and write down those numbers,
even if they are not the final result.
That way we know that you know how to do a table look-up
without the help of a calculator.
|
| Help |
Q:
I cannot get a tutor in my time range so I was wondering
if I could possibly meet with you and get some help?
A:
You are welcome to come to this instructor's office hours.
Remember that the Math 147/148 help room (WH 234)
is open most of the time on weekdays,
probably starting with week 2 of the semester.
This is also the place where your TA holds his/her office hours:
Check the
Helproom 234 schedule
.
I understand that you can ask anyone manning the helproom for assistance.
|
| Help |
Q:
I wasn't able to make my statistics class this morning.
Can I have a brief summary of the topics we went over today?
A:
There are more than 200 students in this class
and close to 25 or even more in your section.
At least one of them should be your best friend,
happy to let you ask them for lecture notes.
|
|
|